NSFW generation is temporarily unavailable. If you don't mind, please submit a "Needs moderator review" report and include a couple of links to LoRAs featuring the character Marinette in the report comments.
🔴 Hi everyone, I wanted to take a moment to talk about the future of my work and this channel. Dedicating so many hours to creating and refining LoRas is becoming increasingly difficult to sustain every single day. If you truly value what I do and want to see more releases, I kindly ask for your support to help keep this passion project alive. If you are in a position to afford it, even a small donation would mean the world to me and make a massive difference.
📌 Support & Subscribe here: https://boosty.to/tetete80099/donate
🔴 Your ongoing contributions directly help cover the rising expenses of this time-consuming hobby. By easing the daily financial pressure, your support allows me to focus my energy entirely on crafting high-quality content, experimenting with new techniques, and giving back to our amazing community without constant worries.
💼 On a side note, I am always open for custom LoRa commissions! I am fully committed to delivering exceptional quality and reliable, tailored service for every project. If you have a specific idea or character in mind, feel free to drop me a DM, and let's bring your vision to life!
I specialize in creating top-tier, highly optimized models for:
· Flux
· Illustrious
· Anima
· Pony
· SD 1.5
· Open to trying other model architectures upon request!
⚡ Flux Architecture
· Pros:
o Extreme Visual Quality: Consistently delivers photorealistic, incredibly crisp, and ultra-high-end images.
o Flawless Character Replication: Offers nearly 100% accurate reproduction of intricate details and original character designs.
· Cons:
o Rigid Prompt Control: Can be quite stubborn and difficult to steer precisely using standard prompts alone.
o No NSFW Capabilities: Completely lacks out-of-the-box support for NSFW content generation.
✨ Illustrious Architecture
· Pros:
o High Image Quality: Yields stunning, highly detailed, and aesthetically rich illustrations.
o Excellent Character Replication: Offers very high fidelity and accuracy in reproducing original character designs.
o Full NSFW Support: Natively optimized for high-quality SFW and NSFW content generation.
· Cons:
o Specific Detail Slip-ups: Can occasionally struggle if the design features complex, highly specific minor details (e.g., precise shapes of necklaces, bracelets, and other accessories).
o Color Layering Issues: In practice, it often struggles to accurately replicate gradient color schemes or colors that blend due to overlapping clothing layers present in the character's original design.
🌌 Anima Architecture
· Pros:
o High Prompt Responsiveness: Flawlessly obeys complex prompts and easily adapts to any tags or descriptions.
o Excellent Quality: Consistently delivers highly detailed, crisp, and visually stunning images.
o Full NSFW Support: Out of the box, it is perfectly optimized for both SFW and NSFW generations.
· Cons (Compared to Illustrious):
o Complex Detail Replication: Slightly falls behind Illustrious when it comes to consistently reproducing intricate minor details in complex, original character outfits.
🦄 Pony Architecture
· Pros:
o Full NSFW Support: Natively built and exceptionally well-optimized for all types of NSFW content generation.
· Cons:
o Average Image Quality: Generates images with modest visual fidelity compared to newer architectures.
o Moderate Character Replication: Offers only average accuracy and consistency when reproducing intricate, original character designs.
o Legacy Hand Issues: Frequently suffers from anatomical hand and finger errors, a flaw characteristic of older image generation models.
🎨 SD 1.5 Architecture
· Pros:
o Massive LoRA Ecosystem: Features an enormous, unparalleled library of existing LoRA models and community assets.
o Full NSFW Support: Offers complete freedom and robust native support for NSFW content generation.
· Cons:
o Poor Image Quality: Produces significantly lower visual fidelity and resolution compared to modern standards.
o Weak Character Replication: Struggles to accurately and consistently reproduce complex original character designs.
o Frequent Deformations: Prone to frequent anatomical errors, causing noticeable distortions in bodies and limbs.
o Difficult Prompt Control: Stubborn and hard to guide precisely, often ignoring nuanced prompt instructions.
Description
Illustrious model for the Marinette Dupain-Cheng Ladybug attire from season 1-5.
Based on 1000 screenshots from the series. Screenshots taken from episode 3 season 2, episodes 8-26 season 3, special episode New York and special episode Shanghai.
Recommended strength 0.6 - 0.7.
I advise you to use built-in triwords for better performance of the model. If consistency with the original is not important to you, delete them.
To add clothes like the original, use the following commands "red suit with black spots all over her body, red mask covering her eyes".
FAQ
Comments (3)
How did you get that?
If you're talking about the model, I took 1000 screenshots from the series, processed them, and uploaded them to CivitAI for training.
If you're talking about the images, I created them on Tensor Art.
Prompt: Super detailed, unbuttoned jacket, white T-shirt, pink jeans, bright bluebell eyes, shoulder-length black hair with blue reflections tied back in pigtails, light pink lips, body like in real life, large pores, fair skin, slender, very little very flat breasts, unreal engine, octane render, droped shadow, bokeh, cinematic lighting, Marinette Dupen-Cheng,
Negative: (low quality, worst quality), displeasing, ugly, poorly drawn, displeasing, simple background, very displeasing, worst quality, bad quality, oldest, deformed limbs, bad anatomy
Base model: IllustriousV1 - Anime 2.5D Style
After which I made some small eye adjustments manually.
@Tetete80099 Thank You!
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.